Dave Weinberger and Linnaeus' paper

I don't know whether it is just me and that people like Dave Weinberger and Clay Shirky are always on the case about metadata, or whether we we are going another round in the process of figuring out what we are actually doing when we try to "organise" information, but right on cue, Dave has a nice, almost elegiac piece about his visit to the Linnaeus Society in London at Linnaeus' paper. He's a good writer and you can bet that he can't possibly leave without finding something there about metadata.

Ms. Douglas takes out a thin pile of 3x5 cards, as soft as handkerchiefs. On each, Linnaeus has recorded in his fine hand one classified species.

This moment, as close as I'll ever get to seeing Linnaeus at work, makes clear how the requirements of the physical world silently persuade us to shape our understanding: Linnaeus' classification resulted from the nature of paper.

Because you only have one card for each species, your order will give each species one and only one place. You will organize them by putting cards near cards like them, naturally producing an ordered series or a set of clusters. As you lay out your cards, like next to like, you are drawing a map of knowledge. (Incidentally, does this raise any issues about IA's using card sorts to underpin data structures? Must bring that up at the next scoff in Sydney)

The largest units are kingdoms, not because Animals, Vegetables and Minerals somehow lord it over the particular creatures they contain but because kingdoms are the most inclusive territories on political maps.

Knowledge thus derives its nature (hmm, structure and boundaries yes, but nature [unless its a Linnaean pun] no) from the paper that expresses it: Bounded, unchanging, the same for all, two-dimensional and thus difficult to represent exceptions and complex overlaps, and all laid out in a glance with no dark corners.

All good, thoughtful stuff and nice to have in the background, but the real information on Linnaean metadata comes just as he steps through the door and finds:

Tucked away near the entrance is a 12-drawer card catalog of the library's books arranged by author. In the work room across the hall sits a computer that allows you to search by abstract, key words, notes, titles, and subjects. There is no hierarchical topic listing.

Exactly.

Comments

[Rick Thomas]

Funny coincidence (two hours before David's post) the way my comment to your previous post makes this exact point. Please see my follow-on comment. Thanks.

[Tom Morris]

Really interesting point actually - I was up at the British Library the other day and all the computers were down. You couldn't do catalogue searches or order books online. You either had to know the exact shelfmark and fill in a paper form to get a request from the stacks or you had to go and look up the shelfmark in these enormous books. This is the same principle: so many books fall under so many categories that such linear organisation is impossible. Research, mine at least, grinded to a halt. It isn't until the systems go down that we realise how damn useful they are.